Analyzing our genome...building on 23andme

[Conductor71]

I am pulling this out of the recent "Possibility" thread because I feel that Paula's original question wasn't given enough consideration. Her suggestion:

Clearly there must be change. So why not anaylze all pwp who desire to - their entire genome? Between the NIH, google, 23andme, MJFF, kinetics, PDF this could be a good way to spend money instead of on mice, lab rats and fruit flies. Wouldn't all our genes thrown into a database identify much about everyone? wouldn't that be worth 100 million dollars instead of yet another 15 yr study that yields nothing?


If I am not mistaken, 23andme analysis is our entire genome and in genetic studies they have proven they are on the cutting edge: they are the first research group to include patients and in just over two years they published a meta analysis the identified two new SNPS associated with PD plus they ensured the results were open source. That is quite a feat compared to snail's pace we are used to. This is all largely the doing of Sergey Brin of Google fame- I think he is one of our unsung heroes. He has given 50 million to research for starters... http://www.wired.com/magazine/2010/0...s_search/all/1

So we have a good solid start in the genetics dept but my question mirrors Paula's. 23andme already has the entire genomes of 6,000 PWP and numerous other GWAS (Genomic Wide Associative Studies) have been published on PD. There is even a database just on PD genomics (www.pdgene.org). So why haven't we gotten any further? Again, it comes down to the profound lack of interest in studying metadata- looking at all studies and finding correlation, IMHO. I know it is all much more complex than this and even if we do find commonalities; this is only one part of the puzzle. Still if someone were to fund what Paula is saying at least we have mad some progress.

We have our genomes, so why not be proactive? Two ways- go public with the data; this can be done on 23andme as well as SNPedia in hope that someone with the interest and ability will look at it and act. Or we can start on a smaller scale here and share only PD related data that in research shows strong association to PD. I am not sure what to do with it though once collected other than the most basic comparison or averaging. However, it seems useful to have this data because we can run queries against them later on.

Anyway, here is a link to the table of the genetic mutations seeming to have the most relevance to PD. Note: only a handful of these are shared currently by 23andme, but you can find more by registering your genome with SNPedia and searching by SNP. It is worth the time and 2.00, I think.

Genome-wide significant summary meta-analysis results of the PDGene database in populations of Caucasian and Asian decent.

We have all this raw data to make use of and maybe this isn't feasible. Any other ideas? It seems we could be learning more from it on our own.

Laura

[johnt]

Thanks for highlighting this Laura.

Apart from giving my sample to 23andme and completing a number of their surveys, I've not looked at the genetic side of PD.

I'm struggling to understand the meaning of the PD meta-analysis table http://www.ncbi.nlm.nih.gov/pmc/arti...-1002548-t002/

So, I'll be grateful if someone would correct my errors as I walk through the first row (GBA) of the table.

Locus/Polymorphism/Location: homes into the location of the genetic difference.

Allele contrast: the actual difference: G where an A would usually be.

MAF = minor allele frequency: 0.01. 1% of people have a G.

N datasets: the number of studies making up the meta-analysis.

N samples: the number of people sampled: 44851

OR: odds ratio: 3.51. A person with this allele is more than 3 times as likely as the average person to get PD.

What I find interesting is the information content in the the knowlege that one has a G rather than an A.

An odds ratio of 3.51 sounds high, but given that on average only 1.6% of people get PD (23andme figure), this equates to a probability of getting PD of 5.6% for people with this mutation.

Or, put another way, given the frequency of 0.01 for this mutation it explains 0.01*0.056*100/1.6 = 0.035% of all PD cases.

This is only one mutation, and others will cumulatively explain more. But it suggests to me that genetics on its own do not explain the aetiology of PD. Genetics are useful, but environment/infection/lifestyle are the key.

Have I got this right?

John

[paula_w]

Unfortunately, i sent two samples and neither yielded any results so i can't be of any assisstance, regarding interpretatin.

paula

[paula_w]

but i thought 23andme tested for only a few genes. LRRK 2 and something else? i know they made a few discoveries,but didn't think they did the whole genome unless you paid for it.

[wordsmithy]

Untrue. They test the whole genome. All free PD members have same benefits as all other members. 23 and me has been so helpful to me! And btw, I am one of the many who do have the GBA allele. I have read that this is turning out to be the most common marker for PD.

[Conductor71]

Originally Posted by wordsmithy

Untrue. They test the whole genome. All free PD members have same benefits as all other members. 23 and me has been so helpful to me! And btw, I am one of the many who do have the GBA allele. I have read that this is turning out to be the most common marker for PD.

Just as Wordsmithy says they test the whole enchilada. Paula, we all carry the LRRK2 gene plus all the others associated with PD. It is a mutation or anomaly in genetic coding that makes us more or less likely to develop PD. The allele is the point at which our code stays on the straight and narrow or makes a turn.

23andme is an awesome source because it is dynamic. They started testing only for LRRK2 but did not stop there. They test us for mutations as they are discovered and then validated. We now have been screened us for approx 8-9 mutations associated with PD including the one linked to alpha-synuclein levels SNCA and MAPT which is misbehaving mitochondria link (if I recall correctly).

I think we might benefit from us sharing at very least whether we are carriers of the PD mutations. For example, some mutations are associated in young onset. If we could pool our data, we might find some links; albeit crude, that has yet to be published.

Worsmithy, do you know the cumulative effect of having 4 alleles that slightly increase PD odds? If they are alone 1.04 increase than that make me 4 times more likely?

Laura

[Conductor71]

John,

You have hit all the important concepts and then added to what little I know with your statistical acumen (I am horrible at maths).

The table is an aggregation of the most replicated genes from genome wide studies that have to date been connected to PD. In other words, there are more than 18 genes now identified. However, up to this point, no one had bothered to analyzed the many genomic studies completed and called out the most commonly occurring genes and polymorphism (mutation). The allele is the interesting point because that is where diverge from everyone else's genetic code. That is the key to finding out whether you have increased chance of disease; you than have "carrier" status. In the long run this may even impact our grandchildren as several genes are auto recessive.

I chose this table because it highlights the mutations that we are all more likely to have. If we were to collect data, this, given the explanation above, seemed like a good place to start. I was motivated by a PWP on 23andme who is collecting mutations related to Vitamin D deficiency and looking for parallels. This is precisely what is not being done (yet) with our data. Why sit around and wait for 23andme to begin running algorithms when we can get a bit of a head start. For example, one of the newly reported mutations at BST1 suggests that calcium levels play a role in things and interestingly enough there is a calcium channel blocker in trials for use in PD. Obviously, I am oversimplifying...

As for risk factor, I think it involves more than we touched on here. Your question is the same I have: what cumulative effect does four different "positive" polymorphisms(AKA SNPs) have? New research highlights that number of mutations within a specific gene might result in early onset PD. I am horrible with statistics, but these gene pools number upwards of 60,000 people. Surely these gene mutations are significant?

While you are correct that genetics plays but one part in sporadic disease, researchers are expressing that genetics plays a much larger role than the 10% they claim as responsible for Early Onset disease. Please see this link to Nature Reviews Neurology 2011 : Parkinson disease: Genetic testing in Parkinson disease—who should be assessed?

Here is link to original article: http://www.ncbi.nlm.nih.gov/pmc/arti...3/?tool=pubmed

Link to genetic info on 23andme: https://www.23andme.com/gen101/graphics/genetics/

Maybe this is way too complex to even think of collecting data but at the very least we could share our PD mutations/SNPs on 23andme in case someone else wants to take action.

HTH

Laura

[EmilyDrabant]

Hello,

My name is Emily Drabant and I manage the Parkinson's Research at 23andMe. I thought I would post a reply to some of the questions that have been posed.

Everyone in our Parkinson’s Community receives the same genetic analysis as our general customers.

Your DNA is made up of several billion letters – A, T, C, and G. All human genomes are mostly identical to each other However, there are small variations scattered throughout. These mostly come in the form of single nucleotide polymorphisms, or SNPs (pronounced “snips”). This just means that in one specific location, I could have a G and you might have a T. Where you have an A, I could have a C.

At 23andMe we analyze your DNA through a process called genotyping. Instead of reading every single letter of your genome in order, the genotyping process just samples individual letters at specific SNP locations. Our chip analyzes about 1 million different SNPs that are spread throughout your entire genome. Therefore, the test is classified as genome-wide. (This is in contrast to full-genome sequencing, in which we would read every single one of the 3 billion letters in your DNA.)

Some of the SNPs we test for have already been linked to health conditions, such as the LRRK2 G2019S variant (linked to PD) or the APOE variations (linked to Alzheimer’s). In your Parkinson’s risk report, we tell you what your genotype is for a set of 10 SNPs that are already known to be associated with PD.

However, we simply don’t know yet know all of the genetic variants associated with Parkinson’s. Therefore our research is focused on finding new associations with other SNPs and Parkinson’s Disease.

Our scientists combine your survey answers with your SNP data and then compare you with everyone else in our database. We look for associations between certain SNPs and certain survey responses, such as whether or not you have Parkinson’s Disease, and if so, was your first symptom a hand tremor or something else?

Ideally, a certain SNP stands out and is strongly associated with a certain survey response. Then we would look at the gene that is linked to that SNP, and start to think about how that gene might play a role in the progression of Parkinson’s.

If you are curious about a certain variant that is not included in our Parkinson’s risk report, you can always check your raw data. Simply log in to your 23andMe account. Then click on Account in the upper right corner of your screen, and then click on Browse Raw Data. You can enter the name of a gene or a SNP ID (in the form of an rs number) into the appropriate boxes. If our chip tests for that area of your genome, then we will provide your results.

If you have more questions about 23andMe in general or our Parkinson’s research in particular, you can email us directly at** You can also look at our Parkinson’s page at **

[paula_w]

i didn't get results because my two spit samples came up with no results. so i don't understand as much as some of the others may. can i ask a few questions?

how did you or whomever find the genes associated with pd? LRRK2 is somewhat cultural is it not?

so to test the entire human being would yield billions of data and that is unreasonable?

or could that be the answer?

thank you,
we want to hear from scientists in the worst way. we know much of it and need the scientists to confirm and /or explain.